Extract objects from images

This code was written to allow objects to be saved as images from larger images. The code creates directories of objects under the output/ directory and saves each image (object) under its corresponding directory. Object detection is performed by YOLOv3 - object detector.

The Code

Prerequisites

  pip install tensorflow==2.0.0
  pip install keras==2.3.0
  pip install pillow

Before executing the main.py file:

• Place your images under input/ directory.
• Download pretrained model file and place it under the model/ directory.
• You can select the objects you want to cut from the images by changing the default value of the objects_to_extract argument found in the main.py file.

After running the program you can find your output under the output/ directory.

License

Distributed under the MIT License. See LICENSE.txt for more information.

Contact

Dudi Biton - dudi709@gmail.com